the social metrics

"Lean analytics" to change business decision making →

analyticisms:

Interesting new concept.

This could signal a shift from broad business analytics to minimalist data decision making and an analytics focus on ‘one key metric’. Interesting take on the lean experimentation approach.

Using Customer, Affiliate and Influencer Data to Geo-Target Your Marketing

Word of mouth and a strong referral network is becoming increasingly important. This has followed the growth of social media, review sites and more connected consumers who can often influence others to buy a product more so than an ad or product offer. 

With that in mind, I’ve been mashing together data about our customers, affiliates and influencers by US state. 

The idea is to model out a potential ‘referrer network’.

That means understanding where we have a good customer base that can refer others to our products. Or a strong affiliate network that resells our products. Or where influential businesses, like bookkeepers, might recommend new customers to our products.

With a bit of modeling, data on these groups can be mashed together to identify where we have the greatest potential for word-of-mouth or affiliate marketing. 

It also highlights locations where we don’t have this support base and, therefore, need to invest more in marketing activities. 

The point of all this is to look at more advanced ways that we can target our marketing spend across the US and be more efficient with how and where we market our products. 

The idea of the referrer network model helps us understand where we already have customers, affiliates and influential professionals recommending our products for us. This means that we can target locations where our referrer network is not as strong and spend marketing dollars on under-served areas. 

I’m still in the early days on this type of mash-up model but I’m starting to identify clear geographical gaps in our referrer network. I’ve mapped out the mash-up using Tableau. States are color-coded to understand where we don’t have a strong referrer network and potentially need to invest in targeted marketing efforts. 

Please note: This is just example data for illustration purposes only. It is not actual company data.

What do you think about this approach?

Sound Measures of Good Community Management « The Community Manager →

My latest tidbits over at The Community Manager that looks across measures of value of community managers and measures of collaboration — offering some ideas on measuring good community management overall. 

Competitive Intelligence and Its Under-Appreciated Role in Social Media Research

I think one of the under-appreciated aspect of social media, when you run an ROI analysis, is social’s ability to provide you with all sorts of timely competitive intelligence. 

Promotions, product issues, advance notice of developments, customer care challenges and share of voice - all of these competitive data points and more can be found through solid social media research. 

Services like Radian6 and NetBase mention the competitive aspect in many of the sales pitches I’ve heard but I rarely see competitive intelligence being given a prominent spot in the social media ROI commentary. It’s usually a side note or a ‘matter of fact’ line in a footnote. However, the potential of social to deliver a higher level of real-time competitive intelligence to layer on your other market research means it deserves a nice spot in the headline. (Of course, that might just be what I read) 

All that transparency means your competitor could also be spying on your organization through the lens of social media. Which might feel a little odd but could also be a great catalyst for innovation as brands compete to outperform each other thanks to newly-found market awareness.

Where Market Mix Modeling Meets Micro Attribution

I saw an interesting tweet the other day from the head of Coke’s research team, Stan Sthanunathan, that said:

Need to balance between macro level mix models and micro level attribution models

The funny thing about that tweet is that I’m going through that exact ‘balancing’ process right now - so the comment reached out from the Twitter stream and grabbed my eye. 

One of the processes we run is market mix modeling in order to understand where our marketing projects impacted sales. The models helps us by telling us the sales contribution in percentages for each channel (e.g. social media). 

Those percentages then help us plan and allocate funds for the next year. 

But models take data, a lot of data. And they take time to run on the data sets so our market mix modeling (MMM) is only run once a year. 

In the interim, during the marketing projects, we try to look at the micro level attribution data to tell us how our specific projects are doing. That means going one step below the channel level and looking at what impact each campaign is having and whether it is meeting the expectations from the overall MMM predictions. 

In social media, for example, we might analyze how a Facebook promotion is doing or how some paid Twitter are performing. 

The key is to triangulate the data between what we expect the channel to be driving and what the micro attribution data is telling us as indicator of marketing efficiency. 

Sometimes that means we rely on proxy data for micro attributions - like engagement or form completions - where the data isn’t financial metrics but we know it has a strong correlation with sales or subscriptions. 

So MMM doesn’t replace micro attribution analyses, or vice versa. They both complement each other. 

The key to balancing these projects is to match MMM, micro attribution and campaign measurements to tie all of the business intelligence to a measurement framework. 

The measurement framework helps tie metrics to business planning to campaign development and hopefully sets expectations across the business. It also ensures the overarching measurement practices are well-thought-through and that the measurement plan hopefully rises above politics or loud-speaking vendors looking for some of the pie.